We are entering the era of artificial intelligence, which promises not only new opportunities and benefits, but also many new challenges.
How is robotization hitting the labor market and is it necessary to introduce a four-day working week because of it?
How are gadgets changing our brains?
Are youth getting dumb?
The director of the Higher School of Mathematics and Applied Physics, Head of the Laboratory of Neural Network Technologies and Artificial Intelligence at Peter the Great St. Petersburg Polytechnic University, Professor Lev Utkin, spoke about the near and distant future that awaits humanity.
Lev Utkin.
Since technological progress is constantly accelerating, there must come a time when people no longer understand anything. By 2030, the computing power of computers will be equal to the human brain. This date is given by the Economic Policy Commission of the US Congress.
Promotional video:
Computers are like children
How long has artificial intelligence appeared and what is hidden behind this term?
- Artificial intelligence has gone through two stages of development. First, the so-called expert systems appeared. In the 1980s, they were considered the most explosive scientific direction. These were systems based on expert opinion. For example, a super professional doctor formulated the rules. Simplified, it looked like this: if a person has a runny nose, high fever and a sore throat, then he has an acute respiratory disease. If at the same time there is also darkening in the lungs, then with a certain probability the patient has pneumonia or tuberculosis. On these rules, according to the principle from general to specific, systems were created that made diagnoses to patients. But the problem is that there are too many situations that happen to patients, and no specialist doctor can describe everything. If you do try to do this, then such systems become very complex. And already in the early 90s they were abandoned due to their unviability. They were replaced by others, which were built on the opposite principle - from the particular to the general. People began to teach computer systems using specific examples.
There is a complete analogy with teaching a child. The kid is shown pictures of cats, dogs, and he gradually begins to imagine these animals. And in general. If he was taught on a striped cat, and then shown a Siamese or a naked Egyptian, he will still say that it is a cat, because in his head he has already generalized its image for himself and can now recognize it regardless of the initial pictures. It's the same with artificial intelligence. The program is "fed" with a bunch of examples, and it itself begins to look for distinctive features, to analyze. For example, if an unmanned car is driving down the street, when it sees a speed limit sign of up to 60 kilometers per hour, it will slow down, because before that, in the process of training, it was shown a bunch of pictures with such signs - including dirty, inverted … And he learned to recognize them.
A new way to fight cancer
You do about the same thing in your laboratory - teach a computer system to recognize cancer in people?
- Yes, we have created a system for automated intelligent diagnostics of lung cancer. Of all the crayfish, this one is the most common. The trouble is that lung cancer cannot be detected in the early stages. One of the signs is said to be coughing. But this is nonsense. A person begins to cough only when the tumor enters the bronchi. Then, one might say, he was lucky to find out about her early. And if it is in other parts of the lung, it is impossible to feel it until the metastases go and the tumor grows to such an extent that the person already begins to suffocate. But then it is already useless to treat it. Therefore, it is necessary with the help of computer screening to find tumors in the early stages, even in the form of small foci. We solve this problem - in fact, we create the same neural network that we train using a bunch of images (made in 3D) of different patients. Anna Aleksandrovna Meldo, a supervisor from the cancer center in Pesochny, helps us with this. We show the computer the images and explain where the cancer is and where it isn't. Having trained, the system is able to examine all suspicious nodules in the patient's tomogram in a matter of seconds and issue its conclusion. A doctor can spend hours on the same job.
Unmanned vehicles are a source of traffic jams
One of the areas of application of artificial intelligence is unmanned vehicles. When will it become reality?
“The artificial intelligence that exists today is enough for the car to recognize signs, situations and road markings on its own. But I can hardly imagine using an unmanned vehicle in real life, especially in our conditions. Snow will fall a couple of times and the markings will no longer be visible. Or a car passing in front of him will spray a road sign. The second problem is how robotic vehicles will behave in extreme situations. With a man it is clear, he begins to twist the steering wheel back and forth. And machines have a tighter algorithm. If there is an accident in front of him on the road, and there is a solid lane on the right, he will never cross it and just get up. And all the others will stand behind him. Self-driving cars are a potential source of traffic congestion. Until now, no one knows how to solve this problem. A possible option is to build separate roads for unmanned vehicles. Traffic jams there will be minimized.
There is one dedicated line for buses, another for robotic vehicles, and a third for cyclists. There is not enough roadway for all the allocated lanes in the city
- Well yes. Or cities for such movement will have to be built initially. This is what happens in China - they build cities from scratch. For example, a new city near Beijing for 100 thousand inhabitants - specialists in artificial intelligence. As in the USSR, whole cities were once built for scientists, and this justified itself, allowed solving the most important problems. This is how the atomic project was successfully implemented. China is also following this path, and that is why it is now ahead of the entire planet.
How robots will change our lives
Could you predict how our life will change in 10 years? Give at least a couple of signs of the near future, besides self-driving cars
- I'm not a visionary, but it is clear that the level of communication will be completely different. There will be some chips instead of phones. For example, built into the palm. Now even small private companies can buy robots for production. And it is clear that soon these robots will be everywhere. A huge number of people will be released. But this does not mean that there will be massive unemployment. Each round of the industrial revolution frees a person more and more from hard work, brings him more free time. Roughly speaking, 100 years ago people worked for 16 hours, without days off at all, they did not see the white light and only rested on holidays. And now they have two days off, vacations, and many work already less than 8 hours. And now there are hundreds of thousands of scientists. There were much fewer of them 100 years ago. It can be imagined that in 10 years even more people will be relieved of hard and routine work,they will have a lot of free time, they will have the opportunity to engage in science and creativity.
The high efficiency of human society has made it possible to free up time for cultural development.
Maybe it will happen. But so far, even in seemingly rather successful Western countries such as Spain or Italy, there is huge unemployment, especially among young people. And this is a real problem for them. What kind of creativity is there …
- So in these two countries you named, they don't really like to work. Shops are open for three hours in the morning and three hours in the evening. And between them - a siesta. I was with my family in Venice. This is horror. During the day everything is closed - nowhere to drink coffee. And this is the norm for them. I think they lived there, in the West, very comfortably for a while, not because of their hard work or high labor productivity, but because of the collapse of the USSR.
A huge country for a song has sold a lot of things to Europe and America, opened huge sales markets for them. What transit flows went through Finland and other nearby countries. How much money was taken to England and Switzerland … And from all this, their standard of living has greatly increased. But they failed to take advantage of all this - and now they are surprised that they have a crisis.
Do you think this is directly related to Russia?
- Sure. And with China, which seemed to be an appendage of Western firms - a source of cheap labor. And now China is already on its own, and Russia seems to be an enemy.
And yet, imagining our current domestic capitalists and our government, I can assume that the release of many people due to robotization will become a huge problem. No one will give these dismissed people a single penny superfluous with the words: "Come on, develop creatively - write books or draw pictures."
- This problem is far-fetched. Look. To do the same work that 100 people are doing now, only two will be needed tomorrow. This means that productivity will increase dramatically. That is, the profit of the owner of the enterprise will also grow sharply. From 1 ruble, he will receive not 2 rubles, but, for example, 10. And if the state normally builds its social policy and does not give a damn about those 98 people who find themselves on the street, then it will increase the tax for this enterprise. And it will redistribute some of the money in favor of the people. In some countries today, as an experiment, they pay their citizens an unconditional basic income (unconditional basic income, AML), regardless of whether they work or not. They simply provide them with money, because they understand that it will be cheaper for the state than to deepen social strife in society. And if this is not done, then, of course, everything will go haywire and people will only get worse.
If people stop working, if machines do everything for them, won't this lead to irreversible consequences for humanity? Maybe people will stop straining and thinking altogether? Why, if machines think for them now?
- Machines cannot think, they solve problems. For example, detecting lung cancer. But this does not mean that we do not need doctors - they just have to "reformat". For example, a doctor used to examine a patient, but now he will look at a computer. Or create databases with snapshots for the computer.
Don't be afraid of gadgets
How do gadgets affect the brain of students? Don't they form clip consciousness? Do not because of them people lose their penchant for analysis, for solving large complex problems?
- Gadgets are useful. They free you from unnecessary work. I remember when I was a student myself, I wrote lectures to the point of insanity. I didn’t remember, I didn’t understand what the teachers told us there - God forbid we had time to write down, outline what they said. And I already passed exams on the basis of notes. Now I am showing the students a presentation. And if I am a normal teacher and my material is well prepared in different media, then I can give them 10 times more knowledge in one lecture than twenty years ago. Previously, in order to find the necessary textbook, you had to go to the library. And there it turned out that he had already been given to someone. There was no question of photocopying at all. I had to spend a lot of time looking for the necessary information. And now a student in a gadget in a second can see what he needs. This is comparable to the advent of a calculator instead of a slide rule, thanks to which we began to count faster and free ourselves up for other important things.
I am sure that the fact that a person does not need to waste time on all sorts of nonsense, his knowledge only becomes deeper.
An unknown artist saw such a Moscow in the XXII century in 1914. The cycle of futuristic postcards was commissioned by the Einem confectionery factory. On Red Square, the noise of wings, the ringing of trams, the horns of cyclists, car sirens, the crackling of engines. Timid pedestrians flee to Execution Ground.
Interesting Facts
The problem of catastrophic forgetting
One of the indicators of the fact that artificial intelligence is still very far from "natural" is the following feature of its memory. If you train a neural network to recognize an object for a long time, then switch it to recognize another, then it will not even remember the original object. For example, if at first the system was taught to distinguish elephants, then they began to train it on monkeys, but then suddenly they again showed an elephant, then artificial intelligence will not recognize it. Because I already forgot what he looks like.
An elephant on the pages of a medieval bestiary, 16th century.
How students have changed over the years
It is believed that young people today are significantly inferior to their peers in recent Soviet times in terms of mental abilities. Professor Lev Utkin does not observe this. In his opinion, among students there is always a certain percentage of intelligent, average, and lazy. And this percentage does not change. Some deterioration in the school preparation of children is noticeable, but young people today lead a healthier lifestyle - they drink and smoke much less than before. Students go in for sports, their leisure time has become more interesting, their horizons are broader. Because today there are more opportunities for this.
Author: Vladlen Chertinov. IA "Petersburg Word"